An Identifiable Double VAE For Disentangled Representations
This work addresses the challenge of disentanglement in unsupervised learning for machine learning researchers, offering an incremental improvement with a new conditional prior method.
The paper tackles the problem of learning disentangled representations in variational autoencoders by proposing a novel VAE-based generative model with theoretical guarantees on identifiability, achieving superior performance compared to state-of-the-art methods on established metrics.
A large part of the literature on learning disentangled representations focuses on variational autoencoders (VAE). Recent developments demonstrate that disentanglement cannot be obtained in a fully unsupervised setting without inductive biases on models and data. However, Khemakhem et al., AISTATS, 2020 suggest that employing a particular form of factorized prior, conditionally dependent on auxiliary variables complementing input observations, can be one such bias, resulting in an identifiable model with guarantees on disentanglement. Working along this line, we propose a novel VAE-based generative model with theoretical guarantees on identifiability. We obtain our conditional prior over the latents by learning an optimal representation, which imposes an additional strength on their regularization. We also extend our method to semi-supervised settings. Experimental results indicate superior performance with respect to state-of-the-art approaches, according to several established metrics proposed in the literature on disentanglement.